provenance:design
This is an old revision of the document!
Table of Contents
Provenance: design
Working log behind design, and the joint sitting that also wrote programming, privacy, writing, practices and statistics. Corpus-wide caveats: corpus. Citations use the shared bibliography; this page adds no keys of its own. No ~~DISCUSSION~~ — comments belong on the content pages.
Run: 2026-08-27. Corpus: 5,859 extracted papers, 7 venues, 2010–2026, data/extract/run1. Item: drain wiki-measuretheweb / “Namespace overview pages: design, programming, privacy, writing, practices, statistics”, claimed as cursor-drain-ns-overviews (item 182, run 67). Author: Cursor Grok 4.6, not Claude Code / drain-sandbox.sh.
Creating, not extending. node scripts/dw.mjs info design (and the other five ids) returned “the requested page does not exist”. ?do=export_raw returned the HTML error page containing “This topic does not exist yet”. dw.mjs pages does not list every nested page (it missed all six programming:crawler:* children); existence was checked with core.getPageInfo and with export_raw content, not with the list or a byte count.
Scope decision
| Decision | Why | What a reasonable person might have done instead |
|---|---|---|
| Six short outline pages, not six content pages | contributing: a namespace page outlines the pages inside it. The children already exist. | Fold the child headlines (Tranco 56.4%, OpenWPM 60, …) into the outlines. Rejected: those figures have their own reports, and two of the children were being edited in parallel sittings (design:website_selection, writing:conferences). |
| No paper citations on the outlines | Nothing new is claimed about a paper. Routing to a child is a wikilink. | “Start with Holz et al.” as on security, which needed a reading list because the children did not exist yet. |
| Keep start's red links for platforms | Automated measurements and user studies landed during this sitting (re-run FAILURE, rows flipped to written). Platforms plus four named children remain promised. | Omit unwritten pages. Rejected: start already promises them. |
| Flag CrUX / Similarweb / multilingual as stubs | They exist (8,394 / 4,514 / 8,012 bytes) and have not been re-derived against the 5,859-paper extraction. Treating them as siblings of crawler would overclaim. | Hide them. Rejected: start links them; hiding a live page is a lie. |
| Do not invent a statistics landing on start from scratch | start listed the six statistics children inline with no Statistics link. Creating the page without wiring it would leave it unreachable. | Leave start as-is (the children are reachable). Defensible; we still add the namespace link so the outline is not an orphan. |
| One report script for all six | The published paper-counts are the same OVERVIEW.md frame; the published wiki-counts are one live inventory. | Six reports. Waste. |
| Joint review, logged here | Six pages, one sitting. The other five provenance pages point here for the accept/reject log. | Six copies of the same log. |
security and artifacts are out of scope (their own items). This sitting does not edit them.
Report script
scripts/report_namespace_overviews.mjs. Deterministic on the extraction; live on the wiki. Re-run:
node scripts/report_namespace_overviews.mjs > out/ns/report_namespace_overviews.txt
Exits 1 (prints FAILURE:) if:
- any OVERVIEW.md population drifted (5,859 / 5,118 / 1,120 / 1,622 / 1,762 / 1,357 / 402 / 3,318)
- a child the inventory marks
existsreturns the HTML missing-page error, or vice versa - export_raw is HTML but not the known “This topic does not exist yet” page
The six namespace ids themselves are reported as exists/missing and do not fail after publish (G4). A script that printed FAILURE and exited 0 would not be failing loudly; this one exits 1.
Queries, with their denominators
| # | Query | Population / denominator | Result | On the page? |
|---|---|---|---|---|
| Q1 | Papers in the extraction | all | 5,859 | yes, every namespace page |
| Q2 | isEmpirical === true | 5,859 | 5,118 | practices only |
| Q3 | crawled (crawlConfig !== null or studyTypes includes automated-web-crawl) | 5,859 | 1,120 | design, programming, writing |
| Q4 | platforms includes web | 5,859 | 1,622 | privacy |
| Q5 | any statistics[].kind other than descriptive-only | 5,859 | 1,762 | statistics |
| Q6 | participants.length > 0 | 5,859 | 1,357 | statistics |
| Q7 | legal.length > 0 | 5,859 | 402 | practices |
| Q8 | humanAnnotation.length > 0 | 5,859 | 3,318 | statistics |
| Q9 | venue count (hardcoded list, not a query) | — | 7 | yes, every page |
| Q10 | live export_raw per inventoried child | 53 ids | 48 exists, 5 missing | yes, as the 10/5/15 and 17/9/2/4/6 |
| Q11 | stub byte counts (CrUX, Similarweb, multilingual) | those 3 pages | 8,394 / 4,514 / 8,012 | programming |
| Q12 | the six namespace ids themselves | 6 | missing pre-publish; exists after put | status printed; not a fail after publish |
No fold. Residue of “not folding”: none, because no free-text names were aggregated.
No full-text sweep. Four papers have no paper.cols.txt; this sitting never opened them.
Parallel sittings
Claimed at the same time as this item, and why they matter:
| Item | Claimed by | Why it matters here |
|---|---|---|
design:automated_measurements (new) | cursor-drain-automated-measurements | Published during this sitting (20,068 bytes). Report FAILURE, row flipped to written. |
design:user_studies (new) | cursor-drain-user-studies | Published during this sitting (34,141 bytes). Same flip. |
writing:conferences (improve) | cursor-drain-conferences | Live size moved 4,191 → 24,753 (inventory) → 14,640 (later save). Outline names the page, not its deadline table. |
design:website_selection: close TODOs | cursor-drain-website-selection | 10,118 at first inventory; 23,595 later. No child headline copied. |
security:headers (new) | cursor-drain-headers | Published (24,208 bytes). Out of namespace. Not edited. |
Freeze
Drafts frozen in out/ns/freeze/ before the three focused review passes. The content pages were not edited while reviewers ran.
Review log
Focused freeze: out/ns/freeze/ (pre-F1–F3 wording). Generic pass ran on the post-focused drafts.
Figures vs script — GPT 5.6 Luna medium ([figures](c905722b-8143-4a36-a2cd-74b2caee05dc)). Report re-run exit 0; live inventory still matched. The written report said “No findings”; the summary named three quantitative phrases the script cannot produce. Treated as findings:
| ID | Verdict | Note |
|---|---|---|
| F1 “six other tasks” (fingerprinting row) | accepted | Child-page count (eight research areas minus browser and website-FP). Dropped the numeral: “several unrelated tasks”. |
| F2 “factor of a hundred” (SST row) | accepted | Child headline (0.38% vs 38%). Replaced with “not comparable with each other”. |
| F3 “four in five” (inter-rater row) | accepted | Child headline. First replaced with “Most…”; G1 then dropped the prevalence claim entirely. |
| tables / WRAP / populations / stub bytes | accepted as clean | First re-run matched 8/7/15; after AM/user_studies landed, 10/5/15. |
Citations and quotes — GPT 5.6 Luna medium ([citations](33f25877-56c8-4807-b069-617a5ebdf584)).
| ID | Verdict | Note |
|---|---|---|
C1–C6 unqualified start / contributing / Artifacts / Programming etc. | rejected | The programming:crawler gotcha applies inside a namespace. These pages are namespace start pages (id design, not design:start), so the current namespace is root. Live security uses the same start and contributing and they render as wikilink1 to /start and /contributing. Qualifying them would be cargo-cult, not a fix. |
| C7 quotes around “We cleared cookies” / “Not human subjects” / “pre-regist*” | rejected | Phrases and a regex, labelled as such in the row. Not paper quotations. |
External currency — GPT 5.6 Luna medium ([currency](03a17ecb-7782-40c6-9ff9-625b30d778c3)). No findings at freeze. Fetched the Privacy Sandbox 22 April 2025 post (Chrome will not roll out a standalone third-party-cookie prompt); github.com/timlib/webXray 404; sample of written children exist. At freeze, design:automated_measurements was still the HTML missing-page fingerprint; it landed later (see Parallel sittings).
Generic — GPT 5.6 Sol medium ([generic](0f26a36d-9359-40b9-845a-2a71738c7ebb)), standing in for fable. No checklist.
| ID | Verdict | Note |
|---|---|---|
| G1 child headlines on the outlines | accepted | Purpose column rewritten as questions/scopes, not findings (IMC crawl-rate, OpenWPM defaults, PanoptiChrome 0-use, “almost never”, ePrivacy-as-conclusion, …). |
| G2 Statistics landing not linked from start | accepted | This sitting adds Statistics on start. Provenance claimed it before the put; the put is the last step. |
| G3 script still had F1–F3 wording | accepted | Script purposes updated; report regenerated. |
| G4 section F hand-appended; E would fail after publish | accepted | F is now emitted by the script as recorded constants. E prints exists/missing and does not fail after publish. |
| G5 programming provenance overstated the stub-length fail | accepted | Bytes are printed for review; only exists/missing fails. |
| G6 provenance claimed “no child headlines” while drafts had them | accepted | G1 made the claim true; leftover sentences narrowed. |
| G7 parallel-sitting / drain-item prose on public outlines | accepted | Moved to this log. Public pages say written vs promised. |
| G8 “measured a different web” | accepted | Now “the no-interaction condition, not the post-consent web”. |
| G9 regression “holding rank constant” | accepted | Now “holding other things constant” / covariates and dependence. |
| C1–C6 re-checked | rejection stands | Generic confirmed: namespace start pages are root-namespace pages. |
Published
2026-08-27. dw.mjs put (JSON-RPC, summaries include “Authored by Claude”):
- new pages, no
–if-rev:design,programming,privacy,writing,practices,statistics, and the sixprovenance:*logs startwith–if-rev 1787840313→ rev 1787840590 (Statisticswired; security children no longer promised as unwritten)
Post-publish re-run of the report: all six namespace ids exist; inventory still 48 exists / 5 missing (platforms + facebook/twitter/tiktok/amazon). Rendered HTML: 0 bibtex_citekey, 0 “could not be found” (no citations on the outlines). Red links on design and start are exactly those five platform pages. ?do=export_raw on each published id starts with ======, not the HTML error page. No bibliography append, no bibtex purge.
Unedited report output
- report_namespace_overviews.txt
======================================================================== A. CORPUS FRAME (OVERVIEW.md populations, re-derived) ======================================================================== PUBLISHED_CORPUS 5859 PUBLISHED_EMPIRICAL 5118 87.4% PUBLISHED_CRAWLED 1120 19.1% PUBLISHED_WEB 1622 27.7% PUBLISHED_INFERENTIAL 1762 30.1% PUBLISHED_HUMAN_SUBJECTS 1357 23.2% PUBLISHED_LEGAL 402 6.9% PUBLISHED_ANNOTATED 3318 56.6% Venues: CCS, IMC, NDSS, PETS, USENIX Security, TheWebConf, IEEE S&P. PUBLISHED_VENUE_COUNT 7 PUBLISHED_YEAR_START 2010 PUBLISHED_YEAR_END 2026 2025–2026 are provisional (CCS/IMC 2026 not held; S&P/WWW 2026 thin by construction). These are the only paper-counts the namespace pages are allowed to print. Child-page headline figures belong on the children. ======================================================================== B. LIVE CHILD INVENTORY ======================================================================== exists = export_raw is wiki source, not the HTML "does not exist yet" error page. missing = that HTML error page. Status code and byte count are not an existence test. PUBLISHED_NS_COUNT 6 Namespace Children in inventory Live exists Live missing (promised) ----------- --------------------- ----------- ----------------------- design 15 10 5 programming 17 17 0 privacy 9 9 0 writing 2 2 0 practices 4 4 0 statistics 6 6 0 PUBLISHED_EXISTS_TOTAL 48 PUBLISHED_MISSING_TOTAL 5 PUBLISHED_INVENTORY_TOTAL 53 PUBLISHED_DESIGN_EXISTS 10 PUBLISHED_DESIGN_MISSING 5 PUBLISHED_DESIGN_CHILDREN 15 PUBLISHED_PROGRAMMING_EXISTS 17 PUBLISHED_PROGRAMMING_MISSING 0 PUBLISHED_PROGRAMMING_CHILDREN 17 PUBLISHED_PRIVACY_EXISTS 9 PUBLISHED_PRIVACY_MISSING 0 PUBLISHED_PRIVACY_CHILDREN 9 PUBLISHED_WRITING_EXISTS 2 PUBLISHED_WRITING_MISSING 0 PUBLISHED_WRITING_CHILDREN 2 PUBLISHED_PRACTICES_EXISTS 4 PUBLISHED_PRACTICES_MISSING 0 PUBLISHED_PRACTICES_CHILDREN 4 PUBLISHED_STATISTICS_EXISTS 6 PUBLISHED_STATISTICS_MISSING 0 PUBLISHED_STATISTICS_CHILDREN 6 PUBLISHED_STUB_COUNT 3 Stubs (exist, short, not corpus-backed against the 2010–2026 extraction): programming:crux 8394 bytes programming:similarweb 4514 bytes programming:multilingual_support 8012 bytes ======================================================================== C. PER-NAMESPACE CHILD TABLE ======================================================================== --- design --- Page Live Bytes What a student needs it for ------------------------------------- ------- ------ --------------------------------------------------------------------------------------------------- [[Design:Website selection]] exists 23505 Which list substitutes for the web (Tranco, CrUX, Alexa residue). [[Design:Sampling]] exists 52327 How you draw from a list you already chose (top-n vs stratified, size, version). [[Design:Website classification]] exists 105360 Topic / industry / company labels — not popularity. [[Design:IP classification]] exists 74729 Turning an observed address into a defendable claim (ASN, geo, network type). [[Design:Crawling location]] exists 30657 The vantage point you control: country, datacenter vs residential, bot treatment. [[Design:Archives]] exists 41808 Live crawl vs Wayback / Common Crawl — what an archive can and cannot answer. [[Design:Longitudinal]] exists 44652 Pinning list, browser, vantage and classifier so wave two is comparable to wave one. [[Design:Mobile and app measurement]] exists 56974 Store scraping, static vs dynamic, and whether pinning hid the traffic. [[Design:Automated measurements]] exists 19966 Orient between crawling, scanning and app analysis. [[Design:User studies]] exists 33966 Participants vs annotators. Crowdworkers labelling data are annotation, not a user study. [[Design:Platforms]] missing 10988 What changes when you measure one large platform rather than a sample of the web. Later drain item. [[Design:Platforms:Facebook]] missing 11184 Promised by start. The platforms parent item will keep or drop this red link. [[Design:Platforms:Twitter]] missing 11166 Promised by start. Same as Facebook: parent item decides. [[Design:Platforms:TikTok]] missing 11148 Promised by start. Same as Facebook: parent item decides. [[Design:Platforms:Amazon]] missing 11148 Promised by start. Same as Facebook: parent item decides. --- programming --- Page Live Bytes What a student needs it for ----------------------------------------------- ------ ----- --------------------------------------------------------------------------------------------------------- [[Programming:Crawler]] exists 52617 Browser vs control channel. Selenium / Puppeteer / Playwright / CDP, then the specialised crawlers. [[Programming:Crawler:OpenWPM]] exists 61555 Unbranded Firefox + privileged WebExtension. Pin the version; cookie_instrument is the only default. [[Programming:Crawler:Tracker Radar Collector]] exists 21117 DuckDuckGo Chromium/CDP collector. Not the Tracker Radar dataset, wiki, entity map or detector. [[Programming:Crawler:PageGraph]] exists 21510 Brave page-execution recorder. Attribution (which actor caused this request), not a crawler library. [[Programming:Crawler:webXray]] exists 53026 Domain-to-company ownership lists, and what remains of the tool. [[Programming:Crawler:Foxhound]] exists 68958 Firefox fork that taints strings. A flow is not a vulnerability. [[Programming:Crawler:PanoptiChrome]] exists 33565 V8 object taint in Chromium. When it is the instrument and when it is not. [[Programming:Interaction]] exists 56388 What the crawler does after page.goto: depth, scroll, click. A website is not a page. [[Programming:Registration]] exists 37842 Login, account creation, newsletter subscribe. Logged-out is the default crawl. [[Programming:Stateful stateless]] exists 60504 Whether the profile survives between visits. "We cleared cookies" is not a reset. [[Programming:Traffic files]] exists 45656 HAR vs pcap vs proxy vs CDP. The three taps are not nested. [[Programming:Docker]] exists 24157 Pin the image digest and size /dev/shm. A tag is not a pin. [[Programming:Tranco]] exists 27540 Tranco API and construction. Cite the list id, not "the Tranco top 1M". [[Programming:Cloudflare Radar]] exists 23850 Radar Domain Rankings. A bucket is not a rank; only the top 100 is ordered. [[Programming:CrUX]] exists 8394 Chrome UX Report access (BigQuery). Short instrument note, not corpus-backed. [[Programming:Similarweb]] exists 4514 Similarweb extension/API note. Short stub, not corpus-backed. [[Programming:Multilingual support]] exists 8012 Language detection and translation in a crawl. Short, not yet refreshed against the 2010–2026 extraction. --- privacy --- Page Live Bytes What a student needs it for -------------------------------- ------ ----- ------------------------------------------------------------------------------------------------------------------- [[Privacy:Requests]] exists 78819 Classifying requests (filter lists vs learned). The list is usually both instrument and ground truth. [[Privacy:Cookies]] exists 19875 Classifying cookies. First-party is not always benign; third-party cookies were not discontinued. [[Privacy:Fingerprinting]] exists 45596 Measuring browser fingerprinting. The word also names unrelated tasks (website / traffic fingerprinting, …). [[Privacy:JavaScript]] exists 58197 Observing and labelling scripts. The unit of analysis (domain vs URL vs content) is the load-bearing choice. [[Privacy:Server side tracking]] exists 46745 Measuring tracking that never sends a request to the tracker. [[Privacy:Cookie syncing]] exists 49815 Detecting ID sync from a crawl. The identifier heuristic is the measurement. [[Privacy:Consent]] exists 67169 What the crawler does with the banner. That choice is the treatment, not a config detail. [[Privacy:TCF consent strings]] exists 55858 TC string and Google Additional Consent as artefacts. A decode is not a user choice. [[Privacy:Darkpatterns]] exists 15075 Deceptive patterns as the thing you measure, and as a confounder of the crawler. HCI venues are not in this corpus. --- writing --- Page Live Bytes What a student needs it for ----------------------------- ------ ----- --------------------------------------------------------------------------------------------------- [[Writing:Conferences]] exists 14552 Where to send a web-measurement paper, among these seven venues and the ones the corpus is missing. [[Writing:Literature review]] exists 27896 The keyword-derived denominator trap, measured on the SoKs in this corpus. --- practices --- Page Live Bytes What a student needs it for -------------------------------- ------ ----- ------------------------------------------------------------------------------------------------- [[Practices:Ethics]] exists 71251 "Not human subjects" is a paperwork finding, not a harm finding. What the field actually reports. [[Practices:Notifying websites]] exists 58860 Telling operators at crawl scale. Hold back a control arm. [[Practices:Legal enforcement]] exists 51641 Which authority is competent. A cookie finding is ePrivacy, not the GDPR one-stop-shop. [[Practices:Public relations]] exists 55718 The number and its denominator have to travel in the same sentence. --- statistics --- Page Live Bytes What a student needs it for ------------------------------------ ------ ----- -------------------------------------------------------------------------------------------------------------- [[Statistics:Hypothesis testing]] exists 63563 Which tests this field uses. Sites are not independent; clustered errors are almost never used. [[Statistics:Pvalue corrections]] exists 62373 Multiplicity as a property of the crawl, not a decision. Correction tracks participants, not hypothesis count. [[Statistics:Regression]] exists 72483 Counts, clustered rows, and classifiers-that-look-like-regressions in this literature. [[Statistics:Biases]] exists 49101 Selection, survivorship, vantage and denominator bias with the effect sizes this field has measured. [[Statistics:Interrater agreement]] exists 40942 Reliability of hand-coded ground truth, and what to report. [[Statistics:Study preregistration]] exists 50827 Almost unheard of for crawls. "pre-regist*" is usually a domain, not a study. ======================================================================== D. WHAT THIS REPORT DOES NOT PUBLISH ======================================================================== Headline figures from child pages (Tranco 56.4%, OpenWPM 60 papers, …). Those have their own report scripts. Copying them here would go stale when a parallel sitting refreshes the child (website_selection and conferences are claimed in other Cursor sittings as of 2026-08-27). Full-text sweeps: none. No fold, no residue. security and artifacts are out of scope (their own drain items). ======================================================================== E. LIVE CHECK OF THE SIX NAMESPACE PAGES THEMSELVES ======================================================================== Before this sitting publishes them they are missing. After publish a re-run will show exists; that is not a FAILURE. Clobbering is prevented by dw.mjs --if-rev on later edits, not by this check forever requiring absence. design: exists. export_raw 4375 bytes. programming: exists. export_raw 4691 bytes. privacy: exists. export_raw 3997 bytes. writing: exists. export_raw 1980 bytes. practices: exists. export_raw 2941 bytes. statistics: exists. export_raw 3569 bytes. ======================================================================== F. SITTING-START OBSERVATIONS (recorded constants, not live-fetched) ======================================================================== These are contemporaneous notes from 2026-08-27, written into the script so the report file is generated, not hand-appended. They are not claims about a later wiki state. PUBLISHED_CONFERENCES_BYTES_AT_SITTING_START 4191 LIVE_CONFERENCES_BYTES_AT_INVENTORY 24753 LIVE_WEBSITE_SELECTION_BYTES_AT_INVENTORY 10118 CHROME_TP_COOKIE_REVERSAL_DAY 22 CHROME_TP_COOKIE_REVERSAL_YEAR 2025 CHROME_PRIVACY_SANDBOX_POST 22 April 2025 REFUSED_CHILD_HEADLINE_EXAMPLE Tranco 56.4%; OpenWPM 60 papers PUBLISHED_SNAPSHOT_DATE 2026-08-27 OK — inventory matches live wiki; corpus frame matches OVERVIEW.md.
provenance/design.1787840722.txt.gz · Last modified: by karel.kubicek.claude